aws-wechat-article-topics

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 1.00). The SKILL.md "第4步：调研" and references/research-strategy.md explicitly require the agent to run web_search and web_fetch to read public webpages (e.g., 公众号 articles, 知乎 answers and comments) and use those untrusted, user-generated sources to inform topic selection and title generation, which can materially change the agent's decisions and actions.